The data scientist
Working with big data doesn’t mean staring at the computer all day, observes Markus Bundschus, Head of Scientific and Business Information Services in Penzberg. Like many other Roche employees, much of Markus’ time is spent brainstorming with teams from various business areas in Roche. These sessions help Markus gain a greater understanding of how mountains of healthcare data can be managed and analyzed to help improve the lives of patients.
Penzberg, located close to Munich in a green section of central Bavaria, is home to more than 5000 employees from both Diagnostics and Pharma. A wing of one building that connects a group of Pharma offices on one side and Diagnostics offices on the other has been dubbed the PHC (Personalized Healthcare) wing. It illustrates the opportunities Penzberg offers for colleagues from Roche’s two divisions to work closely together.
We have to judge whether or not this data has value for Roche.
Markus had not targeted a career revolving around data and computers, originally intending to study molecular biology. But he became interested in a new bioinformatics program jointly offered by the Technical University of Munich and by the Ludwig Maximilians University, launched at about the same time as the sequencing of the human genome.
Markus recalls: “The professors at the universities looked at what was happening and realized, wow, there is so much data around, who is going to analyze that? How are we going to gain new insights from all this information?” His studies encompassed molecular biology, computer science, mathematics and statistics—a wide range of topics but all essential to interpreting data.
After finishing his studies, Markus earned his PhD and worked for Siemens at their offices in Munich and Malvern, Pennsylvania, a Philadelphia suburb. He worked on projects involving an artificial intelligence tool with potential for predicting new cancer genes. This experience was a perfect fit to pursue a career with Roche that began four years ago. At Roche, Markus also took on his first leadership role.
Markus observes that it can sometimes be a bit complicated to explain what his job entails on a daily basis.
“I tend to use the word value chain—at the beginning, we provide contact,” he says. “We look at the interesting and relevant databases, journals and other internal and external sources of data. We have to judge whether or not this data has value for people at Roche. That’s the first part of the job.”
Markus adds: “The second part is information retrieval—filtering the data so that it becomes relevant to Roche. For example, we look for data sources that cover the most interesting genomic cancer experiments published so far. The third part is data analytics—this is about making the data we’ve found ‘actionable’—in other words, it’s data that somebody at Roche can tap into easily.”
Despite the array of tools and programs available, keeping up with technology—not to mention the onslaught of new data—is a never-ending challenge. On any given day, more than 300 new life science patents are filed and databases like PubMed publish about 2400 new articles. And the technology to access this information is changing rapidly as well. For Markus, keeping up with the technological Joneses is “both the hardest part and the most fun part” of his job.
Another challenge is ensuring that the data and the analyses are what the team’s “internal customers”—Roche researchers and product developers, for example—really want. Regular consulting with colleagues across the organization plays a big role in helping the team deliver the right data and analytics.
“We work closely together because from these brainstorming sessions we can better understand what the people need,” Markus says. “These meetings are very beneficial, but sometimes—and I’m sure this is a very common feeling!—I wish there weren’t quite so many.”
For anyone interested in a career as a data scientist, Markus offers some advice: take the time to learn what kinds of data scientists need. Be open to using new tools and to working with teams around the world—for example, colleagues from pRED Informatics.
“Be curious,” he says. “Have an interest in molecular biology. Make the effort to learn the business. If you don’t understand the company’s needs and how your work relates to them, then what you’re doing is your own computer play and nobody will use that.